A Brain Computer Interface (BCI) is the generic term used to describe any kind of system that serves as a communication bridge between the brain (human or not) and an artificial module. It’s a field of research in which wide investment has been made since the 1970’s, especially in the clinical fields and systems’ enhancement. Generally speaking, any kind of brain activity that can be recorded can be used as a means of communicating with another system. Through the use of statistical classification techniques it’s possible to associate certain states or characteristics of the recorded signal – which the experiment subject learns to control - to any procedure, usually mediated by a computer.
Different ways of measuring the brain and kinds of BCIs
Throughout the evolution of Cognitive Neuroscience, many techniques have been developed to help us look and better understand the way the brain works. They range from imaging techniques (like MRI, fMRI, fNIRS or PET), to electrophysiological ones (like EEG, EcG or MEG). While the first category is usually used to obtain high resolution images of brain structures and the second one to register and analyze the electrical activity produced by the brain, with a high temporal resolution – which is why they are the ones mainly used in the field of BCI’s.
As mentioned above, the registering of the electroencephalographic (EEG) activity in a non-invasive way allows us to peak the brain functioning with a high temporal resolution – furthermore, it is now well established that different brain states produce distinct observable activity. With the help of electrodes placed on the scalp, it is possible to feed this activity and their respective variations and patterns to any system capable of classifying and detecting them in real time and act accordingly (making this a field highly interconnected to that of machine learning).
Following closely the developments in signal processing and classification, along with the increasing computational power available, the field of BCI’s was firstly researched as a communication means (for people unable to move, for instance) through the detection of ERPs – event related potentials, small variations of amplitude associated to the presentation of certain stimuli - as well as a way of automatically detecting epileptic seizures. Also, much owing to the first and major financers of such research, the DARPA, the use of BCIs has been always closed associated to the military field, mainly regarding the detection of mental states of fatigue and attention variations, which has led to the development of informatics systems capable of adapting to the mental state of the user.
Currently we have available a considerable range of both research and commercial applications of EEG based BCI systems, and such field is due to receive increased attention in the next through the developing of increasingly efficient classification algorithms and computer power and the cognitive augmentation it might bring.
Although the EEG has been the main technique used for the development of such systems, it has been shown to be possible to integrate electronic controllers directly in the functioning of single cells or even networks. The permanent implant of devices for interpretation and regulation of cortical activity has also been demonstrated.
This has led to a renewed interest in the field and the exploration of new hypothesis, like drug rehabilitation through the detection of relevant cues and stimulation of the brain reward system, rehabilitation after strokes or lesion and even direct transmission of patterns of thought between subjects.
- Tech Summary: Brain-Computer Interfaces
- ThinkTech A blog dedicated to BCI developements
- Commercial EEG BCI System example
- Paralyzed patient controls robot arm using BCI Article from KurzweilAI
- Demonstration of paralyzed patient using robot arm from Nature Magazine YouTube
- Demonstration of a blind patient with a Retinal Implant reading from Discovery Magazine YouTube
Further Reading & Relevant References
- Anderson, J. (1980). Neurocomputing. Cambridge: The MIT Press
- Muller, D. (1995). Towards brain–computer interfacing. MIT Press, Cambridge, MA, 409–422.
- Niedermeyer, E., & Lopes da Silva, F. (2004). Electroencephalography: Basic Principles. Clinical Applications and Related Fields. London
- Vidal, J. (1977). Real-Time Detection of Brain Events in EEG. IEEE Proceedings, 65 (5), 633–641
- Parasuraman, R. (2003). Neuroergonomics: Research and practice. Theoretical Issues in Ergonomics Science, 4, 5–20.